<span style="font-weight: bold;">Fall Meeting Friday, October 26th, 2007 1 pm - 4 pm.</span><br style="font-weight: bold;"><br>The Fall Chapter will be held at the Anshutz Medical Center (formerly Fitzsimons Hospital. ) next Friday, October 26th from 1 to 4 pm. This is the first time we have held a meeting at this location. There are several reasons for holding the meeting at this location. Many of our members are in the biostatistics, bioinformatics and health science related fields. After reading and hearing about many groups moving to the new location over the past decade, this might the first chance some of us have had to visit the new facilities.
<br><br>The talks described below cover a variety of topics both in and outside of the health sciences field. As always, we hope that these talks will appeal to both those involved in the field as well as those outside. A tentative agenda follows.
<br><br>Starting at 1pm.<br><br>Laura Saba - University of Colorado Denver and Health Science Center<br>Calvin Croy - University of Colorado at Denver and Health Sciences Center<br>Manuel Lladser - University of Colorado - Boulder
<br><br>Break with refreshments ~ 20 minutes<br><br>Dennis Lezotte - University of Colorado Health Sciences Center<br>Derek Sondregger - Colorado State University<br><br>With the exception of Denny Lezotte's talk which is 45 minutes, the talks will be approximately 25 minutes. With a little drift, we will conclude at 4pm. Abstracts are below.
<br><br><br><span style="font-weight: bold;">Laura Saba<br></span>Department of Pharmacology, University of Colorado Denver and Health Science Center<br><br><span style="font-style: italic;">Why Mice Drink: Identifying candidate genes from microarray analyses
</span><br style="font-style: italic;"><br>Often differential gene expression analyses result in a large list (>100) of potential candidate genes for a disease or condition. We have developed several filters, both statistical and biological, to help distinguish which genes have the highest potential of clinical significance. This series of analyses is illustrated using data from a mouse model of alcoholism. Data and several statistical tools are publicly available at
<a href="http://phenogen.uchsc.edu">http://phenogen.uchsc.edu</a>.<br><br><span style="font-weight: bold;">Calvin Croy</span><br style="font-weight: bold;"> <br>American Indian and Alaska Native Programs<br>
Department of Psychiatry<br>University of Colorado at Denver and Health Sciences Center<br><br><span style="font-style: italic;">A Brief Introduction to Multiple Imputation</span><br style="font-style: italic;"><br> Popular methods for dealing with missing data in statistical analyses are to delete all observations with missing values or to replace each missing value with a single estimated value. In this talk I will present a high-level introduction to multiple imputation (MI), a less common but often superior method for handling missing data. In multiple imputation several copies of the dataset are created that contain different estimates of the missing data values. The statistic of interest is computed separately from each copy dataset and then combined into a final single value and an associated standard error is also derived.
<br><br>My talk will begin with a conceptual overview of MI and its merits over case/listwise deletion and single imputation. Next I will address the key Missing at Random assumption and how it differs from the similar but unnecessary Missing Completely At Random assumption. I'll also show in general how the final statistic of interest and its standard error are computed. Finally I'll conclude by mentioning software for conducting MI, touch upon best practices, and share a variety of online resources for MI.
<br><br><br><span style="font-weight: bold;">Manuel Lladser</span><br style="font-weight: bold;">The University of Colorado Department of Applied Mathematics<br><br><span style="font-style: italic;">TBA</span><br style="font-style: italic;">
<br><span style="font-weight: bold;">Dennis Lezotte </span><br style="font-weight: bold;">Medical Informatics and Biostatistics<br>Director, Bioinformatics PhD Program<br><br><span style="font-style: italic;">The Colorado School of Public Health Consortium - More than a Thought!
</span><br style="font-style: italic;"><br>Over the past two years, many Coloradans including legislators, public health practitioners, academic leaders, educators and non-faculty partners, have worked tirelessly to conceive of and plan for a School of Public Health (CSPH) that serves Colorado and the Rocky Mountain region. Members of the Colorado School of Public Health Initiative have successfully navigated many obstacles to make the dream of a CSPH a reality. However, before June 1, 2008 many additional operational, logistical and policy decisions must be made and implemented in order to begin the final accreditation phase for this new and exciting initiative.
<br><br>The CSPH is a true consortium consisting of several academic institutions, including programs at CSU, UNC, UC Denver and UCHSC. The CSPH will be an accredited school of public health, providing MPH degrees in the five essential areas of public health concentrations, including: Biostatistics, Epidemiology, Behavioral Health, Environmental Health and Health Systems and Policy. In addition, the CSPH must have established PhD programs in three of these? areas. Currently, a DrPH is being developed as the third
<br>PhD program to compliment the two existing PhD programs in Biostatistics and Epidemiology at the HSC campus.<br><br>The purpose of this presentation is to describe the proposed strategies for providing a multi-campus MPH/Biostatistics degree emerging from the planning efforts of many faculty representing the consortium schools. The mission, vision and values statements have been articulated and initial curricula and program logistical and operational strategies are being
<br>solidified for implementation in 2008. This presentation will describe the consortium MPH/Biostatistics degree program, including anticipated teaching collaborations, student and resource sharing opportunities, the single degree concept and multi-institutional faculty governance issues being anticipated. In addition, the MPH/Biostatistics degree program will be put in context with existing degree programs currently in place on each of the
<br>participating campuses. Our goal is to demonstrate that Statistics, Biostatistics, and Public Health Statistics are alive and thriving in Colorado.<br><br><span style="font-weight: bold;">Derek Sonderegger </span><br style="font-weight: bold;">
Colorado State University<br><br><span style="font-style: italic;">Using SiZer to detect thresholds in ecological data.</span><br style="font-style: italic;"><br>Ecological
systems can display substantial changes in a state variable due to
small changes in environmental conditions. Such changes are
characterized by a nonlinear relationship between the value of the
state variable and one or more environmental variables. Documenting the
magnitude of change and environmental conditions that give rise to
threshold responses is important to both the scientific community and
the agencies charged with ecosystem management. A threshold is defined
as a significant change in a response variable given a marginal change
in environmental conditions. In this paper we demonstrate the
usefulness of a derivative-based method for detecting ecological
thresholds along a single covariate. The Significant Zero crossings
(SiZer) approach employs a nonparametric method to approximate the
response function and its derivatives and then examines the how those
functions change across the range of the covariate. SiZer makes fewer
assumptions than conventional threshold models explores a full range of
smoothing functions. We believe SiZer is a useful technique for the
exploratory analysis of many ecological data sets.<br><br><br><span style="font-weight: bold;">Directions and Parking</span><br style="font-weight: bold;"><br>Directions and Parking Information can be found at <a href="http://www.uchsc.edu/fitzsimons/maps">
http://www.uchsc.edu/fitzsimons/maps</a> .<br>The new center is located approximately at Colfax and Peoria Street - in Aurora.<br><br>On the Anschutz Medical Campus, there are daily cash customer parking lots for UCDHSC Patients and Visitors:
<br><br>Ignacio Lot (511) - located in back of the Administration Building (Building 500) on East 19th Avenue. Neon orange signs posted on campus will also direct you to the UCDHSC visitor parking lot. There are two pay and display machines that are available to render payment. These machines will accept bill denominations up to $20.00 and coins. The flat $4 rate has been change to the following:
<br> 1 hour or less - $1 <br> 1 – 3 hours - $2 <br> Over 3 hours - $4 <br>After 5pm and weekends - $1<br><br>Lot 504 - located on the west side of Uvalda Ct. between 17th Place and 19th Avenue is a metered lot for short term visitors. The rate is $1 per hour. This lot was opened to replace the parking meters that used to be along 17th Place in front of Building 500.
<br>